Optimized sequence
Codon-optimizes a peptide or coding sequence. Every codon stays editable by hand. Optionally it also designs alternative sequences for the same peptide.
Auto treats the input as DNA when it is only A, C, G, T and a whole number of codons. Force it when a peptide happens to look like DNA, such as GATTACA.
How hard to steer away from CpG dinucleotides, which are methylation targets and are associated with transgene silencing. On a 450 aa protein: off gives ~4.5 % CpG, mild ~2.5 %, strong ~0.7 %. Strong costs little in codon frequency but lowers GC by a couple of points.
Codons used less often than this share of their amino acid are never chosen. Raise it to stay with common codons, lower it for more variety when many variants are wanted. An amino acid always keeps at least its top two codons.
Runs of this many identical bases get flagged. Long runs cause polymerase slippage and are a common reason synthesis vendors reject a construct.
The length of identical stretch counted as a repeat. Variants are built to share nothing this long with each other or with the optimized sequence, and internal repeats of this length are reported. 12 to 15 suits short elements, 16 to 20 whole coding sequences.
How many candidate sequences are drawn before the best is kept. Runtime scales with this; 300 is ample for short elements, and raising it helps mainly when the repeat window is tight.
Compare reverse complements too when looking for repeats and restriction sites. Worth leaving on: inverted repeats recombine, and Type IIS sites such as BsaI are not palindromic.
Sequences already in your plasmid that the output should stay different from — an existing P2A, another copy of the gene, a linker. Anything shorter than the repeat window is ignored.